on Martingales

18.175: Lecture 27
More on martingales
Scott Sheffield
MIT
18.175 Lecture 27
1
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
2
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
3
Recall: conditional expectation
�
I
Say we’re given a probability space (Ω, F0 , P) and a σ-field
F ⊂ F0 and a random variable X measurable w.r.t. F0 , with
E |X | < ∞. The conditional expectation of X given F is a
new random variable, which we can denote by Y = E (X |F).
�
I
We require
� that Y is
� F measurable and that for all A in F,
we have A XdP = A YdP.
�
I
Any Y satisfying these properties is called a version of
E (X |F).
�
I
Theorem: Up to redefinition on a measure zero set, the
random variable E (X |F) exists and is unique.
�
I
This follows from Radon-Nikodym theorem.
18.175 Lecture 27
4
Conditional expectation observations
�
I
Linearity: E (aX + Y |F) = aE (X |F) + E (Y |F).
�
I
If X ≤ Y then E (E |F) ≤ E (Y |F).
�
I
If Xn ≥ 0 and Xn ↑ X with EX < ∞, then E (Xn |F) ↑ E (X |F)
(by dominated convergence).
If F1 ⊂ F2 then
�
I
�
�
E (E (X |F1 )|F2 ) = E (X |F1 ).
E (E (X |F2 )|F1 ) = E (X |F1 ).
�
I
Second is kind of interesting: says, after I learn F1 , my best
guess of what my best guess for X will be after learning F2 is
simply my current best guess for X .
�
I
Deduce that E (X |Fi ) is a martingale if Fi is an increasing
sequence of σ-algebras and E (|X |) < ∞.
18.175 Lecture 27
5
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
6
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
7
Martingales
�
I
Let Fn be increasing sequence of σ-fields (called a filtration).
�
I
A sequence Xn is adapted to Fn if Xn ∈ Fn for all n. If Xn is
an adapted sequence (with E |Xn | < ∞) then it is called a
martingale if
E (Xn+1 |Fn ) = Xn
for all n. It’s a supermartingale (resp., submartingale) if
same thing holds with = replaced by ≤ (resp., ≥).
18.175 Lecture 27
8
Martingale observations
�
I
Claim: If Xn is a supermartingale then for n > m we have
E (Xn |Fm ) ≤ Xm .
�
I
Proof idea: Follows if n = m + 1 by definition; take
n = m + k and use induction on k.
�
I
Similar result holds for submartingales. Also, if Xn is a
martingale and n > m then E (Xn |Fm ) = Xm .
�
I
Claim: if Xn is a martingale w.r.t. Fn and φ is convex with
E |φ(Xn )| < ∞ then φ(Xn ) is a submartingale.
�
I
Proof idea: Immediate from Jensen’s inequality and
martingale definition.
�
I
Example: take φ(x) = max{x, 0}.
18.175 Lecture 27
9
Predictable sequence
�
I
Call Hn predictable if each H + n is Fn−1 measurable.
�
I
Maybe Hn represents amount of shares of asset investor has at
nth stage.
�
Write (H · X )n = nm=1 Hm (Xm − Xm−1 ).
�
I
�
I
Observe: If Xn is a supermartingale and the Hn ≥ 0 are
bounded, then (H · X )n is a supermartingale.
�
I
Example: take Hn = 1N≥n for stopping time N.
18.175 Lecture 27
10
Two big results
�
I
Optional stopping theorem: Can’t make money in
expectation by timing sale of asset whose price is non-negative
martingale.
�
I
Proof: Just a special case of statement about (H · X ) if
stopping time is bounded.
�
I
Martingale convergence: A non-negative martingale almost
surely has a limit.
�
I
Idea of proof: Count upcrossings (times martingale crosses a
fixed interval) and devise gambling strategy that makes lots of
money if the number of these is not a.s. finite. Basically, you
buy every time price gets below the interval, sell each time it
gets above.
�
I
Stronger convergence statement: If Xn is a submartingale
with sup EXn+ < ∞ then as n → ∞, X+ n converges a.s. to a
limit X with E |X | < ∞.
18.175 Lecture 27
11
Other statements
�
I
If Xn is a supermartingale then as n → ∞, Xn → X a.s. and
EX ≤ EX0 .
�
I
Proof: Yn = −Xn ≤ 0 is a submartingale with EY + = 0.
Since EX0 ≥ EXn , inequality follows from Fatou’s lemma.
�
I
Doob’s decomposition: Any submartingale Xn can be
written in a unique way as Xn = Mn + An where Mn is a
martingale and An is a predictable increasing sequence with
A0 = 0.
�
I
Proof idea: Just let Mn be sum of “surprises” (i.e., the
values Xn − E (Xn |Fn−1 )).
�
I
A martingale with bounded increments a.s. either converges to
limit or oscillates between ±∞. That is, a.s. either
lim Xn < ∞ exists or lim sup Xn = +∞ and lim inf Xn = −∞.
18.175 Lecture 27
12
Problems
�
I
How many primary candidates does one expect to ever exceed
20 percent on Intrade? (Asked by Aldous.)
�
I
Compute probability of having a martingale price reach a
before b if martingale prices vary continuously.
�
I
Polya’s urn: r red and g green balls. Repeatedly sample
randomly and add extra ball of sampled color. Ratio of red to
green is martingale, hence a.s. converges to limit.
18.175 Lecture 27
13
Wald
�
I
Wald’s equation: Let Xi be i.i.d. with E |Xi | < ∞. If N is a
stopping time with EN < ∞ then ESN = EX1 EN.
�
I
Wald’s second equation: Let Xi be i.i.d. with E |Xi | = 0 and
EXi2 = σ 2 < ∞. If N is a stopping time with EN < ∞ then
ESN = σ 2 EN.
18.175 Lecture 27
14
Wald applications to SRW
I
�
S0 = a ∈ Z and at each time step Sj independently changes
by ±1 according to a fair coin toss. Fix A ∈ Z and let
N = inf{k : Sk ∈ {0, A}. What is ESN ?
I
�
What is EN?
18.175 Lecture 27
15
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
16
Outline
Conditional expectation
Martingales
Arcsin law, other SRW stories
18.175 Lecture 27
17
Reflection principle
�
I
How many walks from (0, x) to (n, y ) that don’t cross the
horizontal axis?
�
I
Try counting walks that do cross by giving bijection to walks
from (0, −x) to (n, y ).
18.175 Lecture 27
18
Ballot Theorem
�
I
Suppose that in election candidate A gets α votes and B gets
β < α votes. What’s probability that A is ahead throughout
the counting?
�
I
Answer: (α − β)/(α + β). Can be proved using reflection
principle.
18.175 Lecture 27
19
Arcsin theorem
�
I
Theorem for last hitting time.
�
I
Theorem for amount of positive positive time.
18.175 Lecture 27
20
MIT OpenCourseWare
http://ocw.mit.edu
18.175 Theory of Probability
Spring 2014
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms .